In our first-in-human phase I studies, ex vivo expanded regulatory T cell (Tregs) transfer reduced but did not eliminate graft-versus-host disease (GVHD). Tregs were not designed to function alone but to cooperate with supporting innate immune cells to function. We hypothesized that myeloid-derived suppressor cells (MDSCs), were the missing cell type required for optimal Treg suppression. Amongst their suppressive mechanisms MDSCs locally deplete essential amino acids placing highly proliferative GVHD-causing T cells under metabolic stress. We showed that IL-13 cultured MDSCs upregulate arginase I, causing in vivo depletion of the essential amino acid, L-arginine. MDSCs cultured with IFNg upregulate indoleamine 2,3-dioxygenase (IDO), which both can augment Treg function and deplete the essential amino acid tryptophan. Due to the relative resistance of Tregs vs Teffectors to amino acid starvation, Tregs would be favored over Teffector cells. We have uncovered that neither MDSC subset alone is completely effective in preventing GVHD. We now attribute these findings to the new observation that MDSCs lose their suppressive capacity upon activation of the inflammasome; a molecular machine that activates proinflammatory cytokine proteolytic processing. Since MDSCs are immature myeloid cells, we propose progression of GVHD injury drives MDSCs into differentiated myeloid cells that lose their suppressor potency. We hypothesize Tregs will promote MDSC suppression by inhibiting inflammasome activation and proinflammatory cytokine driven MDSC differentiation. Because ex vivo expanded Tregs can be converted to Teffectors under inflammatory conditions, we propose that MDSC suppression of inflammation will support Treg function and stability. In aim 1 we will test the hypothesis that Teffector suppression by donor MDSC infusion is subverted by inflammasome activation, resulting in MDSC differentiation and loss of suppression. We will identify the inflammasome components in MDSC-IL13 and MDSC-IFNg cells and inflammasome inciting stimuli and proinflammatory cytokines driving differentiation of MDSCs into nonsuppressive or allostimulatory cells. In aim 2, we will test the hypothesis that essential amino acid catabolic enzymes in donor MDSCs drive Teffectors into catastrophic metabolic stress, augmenting the function and stability of adoptively transferred Tregs, and cooperating with Tregs to maximally suppress GVHD. Further we will test the mechanisms of MDSC-mediated Treg function and stability, while inflammation and inflammasome activation in MDSCs is subdued by Tregs. With the optimal MDSC and Treg therapy, we will test whether the graft-versus-leukemia (GVL) can be retained. When complete, we will have developed new therapies to prevent GVHD and maintain GVL.